One-Slide Calibration of Pathology Foundation Models

Ming Ren Hou; Tianyi Huang

cs.CVOct 6, 2026

One-Slide Calibration of Pathology Foundation Models

Authors: Ming Ren Hou, Tianyi Huang

Organizations: App-In Club · Ryquo

Abstract

Scanner variation changes how pathology foundation models represent the same tissue. We introduce SlideRuler, which uses regions within a slide as internal controls to estimate and correct acquisition-induced shifts in other regions. A transfer map learned from paired rescans enables calibration from a single scan at inference while keeping the foundation model fixed. Across two encoders and five SCORPION scanners, learned transfer reduces mean target-to-source embedding distance by 16.3-38.5% relative to raw embeddings. Comparisons with unrelated same-scanner controls reveal a positive same-slide contribution across all four evaluation settings, including scanner holdout. A source-anchored variant reduces source-feature displacement by 47.7-83.6% relative to learned transfer while retaining most of its alignment gain. By drawing calibration information from the slide itself, SlideRuler offers a path toward more consistent use of frozen pathology models across imaging systems.

Figures & tables

Figure 1: One-slide calibration with learned transfer. Top: paired rescans define the acquisition basis $U$ from residuals centered across scanners within each registered region and train the transfer map. Bottom: a single scan supplies control displacements relative to source prototypes; learned transfer corrects the query while preserving $Qz_{j}$ . Controls exclude the query; training prototype matches exclude its entire slide. The nine-control SCORPION example is schematic.

Figure 2: Slide-specific acquisition information across scanners. Same-slide distance advantage over unrelated same-scanner controls, disaggregated by target device. Positive values favor same-slide controls. Points and bars show mean differences and pointwise 95% physical-slide bootstrap intervals over 48 slides, after averaging 32 within-fold donor permutations. The query embedding is unchanged. Limits are shared within encoder rows; distances use each encoder’s source-training scale.

Encoder	Protocol	Raw	Scanner mean	Affine ridge	Shrinkage	Learned	Anchored
Prost40M	Calibrated	18.654	11.328	9.640	13.181	11.470	12.030
	Holdout	18.654	15.940	13.294	14.801	13.871	14.196
DistillPath	Calibrated	17.551	13.805	14.447	13.808	13.843	14.591
	Holdout	17.551	15.840	15.420	14.690	14.682	14.973

Table 1: Paired embedding alignment on SCORPION. Mean target-to-source distance; lower is better. All corrections are ungated. Bold denotes the lowest point estimate in each row. Scanner mean falls back to pooled mean for a held-out scanner. Distances use each encoder’s source-training scale.

Figure 3: Alignment and source displacement form distinct evaluation axes. Large markers are method means; faint blue/orange points show learned/anchored results for each of 48 physical slides, joined within slide. Lower values on both axes are preferable. Target values average ten regions and four target scanners; source values average ten regions and one fitted model per slide when calibrated, or four models under scanner holdout. Axis limits are shared within encoder rows; distances use source-standardized coordinates.

Figure 4: Control availability across sampling policies. (a–c) Empirical cumulative distributions of benign-gland counts for 35 reference and 34 dense scans per scanner. The horizontal coordinate is $\log_{2}(n+1)$ , with ticks labeled by count $n$ ; dashed lines mark 16 fields. (d) Availability requires at least 16 benign-gland and 16 stromal fields per scan. Columns represent all 35 slides, with unavailable preparations marked. G/H/L denote Grundium/Hamamatsu/Leica. Sampling locations are non-nested.

Reference class	Fields	Patients	Precision	Recall
Benign gland	337	21	89.37 [80.05, 95.72]	89.51 [82.19, 94.40]
Residual stroma	7,028	32	99.98 [99.93, 100.00]	98.16 [96.78, 99.18]
Tumor	960	20	96.32 [91.85, 98.69]	81.40 [71.34, 89.38]

Table 2: External patient-held-out tissue selection. Precision and recall are percentages with 95% patient bootstrap intervals. Counts denote reference-class fields and patients containing that class. Residual stroma is the dataset’s postprocessed remaining-tissue class.

Policy	Grundium	Hamamatsu	Leica	Overall
Reference (512 fields)	12/35	3/35	14/35	29/105 (27.6%)
Dense (2,048 fields)	21/35	17/35	20/35	58/105 (55.2%)

Table 3: PAR control-count availability. Scans with at least 16 selected benign-gland and 16 stromal fields. Each scanner has 35 assigned scans. The three unavailable dense scans remain in the denominator.

Appendix figures & tables11 assets

Supplementary material from the paper’s appendix.

Appendix

Encoder	Protocol	Pooled mean	Direct controls
Prost40M	Calibrated	14.131	12.215
	Holdout	15.940	13.837
DistillPath	Calibrated	14.985	14.051
	Holdout	15.840	14.770

Appendix

Table 4: Additional baseline distances, completing Table 1 .

Encoder	Protocol	Versus affine ridge	Versus scanner mean	Versus shrinkage
Prost40M	Calibrated	$-1.830$ [ $-2.007,-1.649$ ]	$-.142$ [ $-.251,-.035$ ]	1.711 [1.565,1.865]
	Holdout	$-.576$ [ $-.739,-.411$ ]	2.069 [1.979,2.158]	.930 [.792,1.082]
DistillPath	Calibrated	.604 [.411,.798]	$-.038$ [ $-.091,.018$ ]	$-.035$ [ $-.121,.052$ ]
	Holdout	.738 [.553,.922]	1.158 [1.101,1.219]	.008 [ $-.064,.079$ ]

Appendix

Table 5: Learned transfer compared with global alternatives. Positive paired distance reductions favor learned transfer. Intervals use 5,000 paired physical-slide bootstrap resamples.

Encoder	Protocol	Learned, 95% CI	Anchored, 95% CI	Learned gain range
Prost40M	Calibrated	.535 [.427,.641]	.249 [.169,.325]	.472–.591
	Holdout	.303 [.238,.369]	.156 [.099,.213]	.261–.348
DistillPath	Calibrated	.330 [.270,.392]	.031 [.015,.047]	.298–.367
	Holdout	.249 [.201,.299]	.058 [.041,.075]	.220–.277

Appendix

Table 6: Same-slide distance advantage over 32 wrong-slide assignments. Permutations are averaged before physical-slide resampling; the gain range describes variation across the 32 assignments.

		Top-1 retrieval accuracy			Distance worsening
Encoder	Protocol	Raw	Learned	Anchored	Learned	Anchored
Prost40M	Calibrated	93.91	99.74	99.84	1.77	1.15
	Holdout	93.91	99.74	99.58	3.07	2.45
DistillPath	Calibrated	99.79	100.00	99.95	1.77	1.35
	Holdout	99.79	99.95	99.90	2.40	1.35

Appendix

Table 7: Top-1 retrieval accuracy (%) and material distance worsening (%). The worsening threshold is defined in Section 4 .

Encoder	Protocol	Target distance, 95% CI	Retrieval change, 95% CI	Worsening change, 95% CI
Prost40M	Calibrated	.560 [.466,.652]	.104 [.000,.260]	$-.625$ [ $-1.094,-.260$ ]
	Holdout	.326 [.264,.389]	$-.156$ [ $-.521,.156$ ]	$-.625$ [ $-1.302,.000$ ]
DistillPath	Calibrated	.749 [.677,.822]	$-.052$ [ $-.156,.000$ ]	$-.417$ [ $-.833,-.052$ ]
	Holdout	.291 [.241,.344]	$-.052$ [ $-.156,.000$ ]	$-1.042$ [ $-1.562,-.573$ ]

Appendix

Table 8: Paired anchored-minus-learned changes. Positive target-distance change means less alignment; negative worsening-rate change is favorable. Rate changes are percentage points.

Figure 5: External tissue-selector validation. Each patient’s complete field set has total weight one. (a) Row-normalized primary out-of-fold selection outcomes, including unassigned fields at the fixed 0.8 cutoff. (b) Precision and recall with 95% intervals from 2,000 patient bootstrap resamples; recall includes unassigned reference fields. (c) Benign precision and recall for the primary partition and all four post-result randomized partitions of the same 32 patients. The four additional partitions assess sensitivity to patient allocation.

Encoder	Protocol	Learned	Anchored	Global-only	Residual	Identity
		Source displacement		Selected models
Prost40M	Calibrated	3.562	1.861	0	6	0
	Holdout	3.306	1.683	0	24	0
DistillPath	Calibrated	2.365	.388	6	0	0
	Holdout	2.248	.555	12	12	0

Appendix

Table 9: Source-feature displacement and model selection. Counts refer to six unique fits per calibrated setting and 24 per holdout setting. Source means average 48 physical slides, each evaluated under one calibrated model or four holdout models.

Policy	Direction	Slides /35	Patients /20	Training patients, folds 0–4
Reference	G $\to$ H	0	0	0, 0, 0, 0, 0
	G $\to$ L	2	2	2, 2, 0, 2, 2
	H $\to$ G	0	0	0, 0, 0, 0, 0
	H $\to$ L	0	0	0, 0, 0, 0, 0
	L $\to$ G	2	2	2, 2, 0, 2, 2
	L $\to$ H	0	0	0, 0, 0, 0, 0

Appendix

Table 10: Direction-specific calibration eligibility. Each fold assigns 16 training and four held-out patients. Eligibility requires both scans’ controls, approved registration, and four matched tumor fields. G: Grundium; H: Hamamatsu; L: Leica.

Policy	Class	Selected fields	With nearby center	Fraction
Reference	Benign gland	1,667	459	27.5%
	Residual stroma	25,729	12,213	47.5%
Dense	Benign gland	6,251	4,016	64.2%
	Residual stroma	100,660	86,688	86.1%

Appendix

Table 11: Spatial clustering of selected controls. A nearby center is another selected center of the same class within 112

\mu

m. The fraction measures spatial clustering of sampled fields.

Regime	Audit	Raw	Corrected	Raw	Corrected	Raw	Corrected	Cov.
		Shift RMSE		Worst QWK		Grade flip
Acquisition-sensitive	7/8	4.014	1.894	.828	.949	.567	.157	.804
Already robust head	7/8	4.031	1.893	.983	.993	.098	.038	.807
Unseen scanner	7/8	4.449	2.513	.790	.932	.617	.185	.753
Unseen, 1.5 $\times$ range	7/8	6.673	5.925	.643	.729	.714	.479	.365
Matched tissue response	7/8	3.504	1.630	.865	.961	.555	.150	.813

Appendix

Table 12: Complete synthetic regimes. Means include all eight seeds; rejection by the outer screen returns the raw embeddings. Cov.: mean correction coverage. Audit: seeds passing the outer screen.

Stress	Ungated worsening	Coverage	Accepted worsening
One-control conflict	.738	.000	–
Concordant control drift	.964	.037	.990
Morphology out of support	.003	.000	–
Novel acquisition direction	.001	.000	–
Insufficient controls	.000	.000	–
25% selector contamination	.001	.415	.000

Appendix

Table 13: Synthetic stress tests in the acquisition-sensitive regime. Material worsening exceeds 2% of the corresponding clean test regime’s median raw non-source shift norm. Accepted worsening is conditional on acceptance and undefined at zero coverage. Means of conditional rates omit seeds with undefined rates.

Explore similar work

HERO: Histology Encoder for Robust Representation in Oncology
Sep 28, 2026Zhi Li, Eghbal Amidi, Yating Cheng +12Pathology Foundation Models Computational Pathology
Caris Life Sciences, Irving, TX, United States
Robustifying pathology foundation models via fine-tuning
Jul 24, 2026Alexandre Filiot, Oskar Thaeter, Benoit Schmauch +1Pathology Foundation Models Medical World Model
1Waiv · Institute of Pathology, Technical University of Munich · School of Computation, Information and Technology, Technical University of Munich
From Patches to Patients: A study of the tile-to-slide performance transferability in Digital Pathology
Jun 9, 2026Sofiène Boutaj, Leo Fillioux, Maria Vakalopoulou +2Wireless Foundation Models Multiple Instance Learning
Université Paris-Saclay, CentraleSupélec, Gustave Roussy, INSERM, IHU PRISM, Cancer Data Science Unit, France · Université Paris-Saclay, CentraleSupélec, MICS Laboratory, France

Sep 28, 2026, cs.CVJ/K move · Enter open · S save

HERO: Histology Encoder for Robust Representation in Oncology

Zhi Li, Eghbal Amidi, Yating Cheng, Tyson Dawson+11

Caris Life Sciences, Irving, TX, United States

Pathology Foundation Models Computational Pathology Medical World Model Wireless Foundation Models Representation Quality

Open paper AskPDF arXiv

Foundation models trained on large pathology image corpora now provide strong, transferable representations for computational pathology. Over the past few years a series of such models has been released, each trained on more slides than the last; on standard classification and segmentation benchmarks, the leading models are now separated by small margins. In clinical use, however, the foundation model is applied to images from hospitals, scanners, and staining protocols outside its training data. Encoders generally embed these acquisition factors alongside biological information, which may introduce downstream errors and hinder safe clinical adoption. A pathology foundation model should therefore be robust to acquisition shift without giving up representation quality, yet robustness is seldom the axis along which models are compared. In this report, we introduce HERO (Histology Encoder for Robust Representation in Oncology), a ViT-G/14 pathology foundation model trained with the DINO and iBOT objectives and refined with high-resolution Gram anchoring on a morphology-balanced corpus of 500 million tiles from approximately 575,000 clinical whole-slide images. Across the evaluated public benchmarks, HERO shows the strongest robustness to center, scanner, and stain variation among the compared state-of-the-art foundation models, performs comparably on tile-level classification, segmentation, and gene-expression prediction, ranks first on average across 39 evaluated slide-level clinical tasks, and, under an equal-weighted framework-level analysis, has the best average rank across the six benchmark frameworks.

Abstract

Explore similar work

HERO: Histology Encoder for Robust Representation in Oncology

Robustifying pathology foundation models via fine-tuning

From Patches to Patients: A study of the tile-to-slide performance transferability in Digital Pathology